OSLER inHealth Statistical Software

OSLER inHealth (Open-Source Learning Environment for Research on Individualized Health) consists of R-based statistical functions that incorporate Bayesian hierarchical models to analyze and visualize complex health data to support health-related decision-making.  

The first tool available in the OSLER inHealth suite of functions is baker: Bayesian Analysis Kit for Etiology Research. baker implements hierarchical Bayesian models to estimate population etiology fractions for multivariate binary data and can be accessed at https://github.com/zhenkewu/baker.

Coming soon:  ProstatePredict, an R package that implements a hierarchical Bayesian latent class model for predicting the underlying prostate cancer status of patients with localized disease, will be released soon. Preliminary versions of the software are available at https://github.com/rycoley/prediction-prostate-surveillance.

OSLER inHealth is also developing an R package that will implement a generalized hierarchical Bayesian latent health state model. Users will be able to upload clinical data from multiple sources and specify relationships between latent states of interest and observed data. This software will accommodate mixed data types, informative missing data patterns, and causal effects of both external covariates and endogenous interventions on health.